Budgeting Wins
I was reading an article by Nate Silver a few weeks ago when I ran across a paragraph that articulated many of the thoughts I’ve been having on how teams should approach their offseason plan. Essentially, Nate put my thoughts into words, and then expressed them in a way that made sense of the jumble I had going on upstairs. Here’s what he wrote under the context of suggesting moves for Walt Jocketty in keeping the Cardinals as contenders in 2007:
Set a wins budget, not a payroll budget. It’s my belief that most major league baseball teams go about their budgeting the wrong way. Sometimes it’s worth spending more than you might have been planning on if you can acquire a player who will get you over the hump and into the playoffs, since the financial rewards for making the postseason are substantial. Other times, your team is so far from making the playoffs-–or such a cinch to make them-–that paying market price for free agent talent just doesn’t make sense.
Put differently, rather than figuring how much money they want to spend, baseball teams should determine roughly how many wins they’ll need to make the playoffs, and spend on payroll until that they have the talent on hand to realistically reach that goal. For the Cardinals, the magic number is probably 90 wins; it would be higher if they played in a tougher division.
This is a wildly different concept than most teams operate under. Because almost every major league team is run as a corporate operation, the baseball operations department is given a budget for player payroll that fits into a fiscal year line-item for the corporation as a whole. Essentially, the accountants tell the front office how much money they can spend and tell them to do as well with that as they can.
While acknowledging that there are some issues with what I’m suggesting, I believe this design for setting payroll is inefficient and could be vastly improved. Instead of trying to cram as many wins as possible into a preset number of dollars spent, teams should figure out how many dollars they need to spend to reach their preset number of wins targeted.
Basically, Nate and I (independently, as we haven’t talked about this) support a 180-degree change in the way the baseball operations departments interact with the organization’s accountants. Rather than handing the baseball ops people a similarly sized check every year, baseball teams would be better off going to requisition-style budgeting concept.
How would this work?
We’ve talked a lot about marginal win values on the blog. Essentially, the concept of marginal wins is that a team of league minimum players would win 50 games or so, and every dollar spent over the league minimum is being used in hopes of gaining wins 51 and up. These wins are the marginal wins, and the excess payroll is the marginal dollars spent. This kind of analysis allows us to look at the league as a hole and say that a marginal win is worth about $2 million in salary.
However, league wide averages of marginal win values start to lose some of their importance when applied specifically to one team. Every team has a different roster of committed salaries and budgets for the following season, and so each team will have a different marginal dollar amount to spend on building their roster. The Yankees can blow the $2 million per win number out of the water and not blink an eye, while if the Marlins tried to spend $2 million per win, the ownership would flip out and demand a firesale.
In a win-budget concept, a team would do a rational, honest evaluation of how many games the current roster could be expected to win if each additional hole was filled with a replacement level player, and then request a player payroll budget based on the amount of marginal wins they’re going to attempt to add that offseason. For a team that believes it has a 60 win team in the organization, spending $30 million to buy another 5-10 wins is pretty foolish. However, if a team has 90 wins sitting around, spending $30 million to make yourself a 95-100 win team is well worth justifying. A team operating on a win budget system would spend significantly more when their team is contending and significantly less when their team is rebuilding.
A move away from a fiscal year budgeting system would give the baseball operations department much more flexibility in their long term planning and allow organizations to allocate resources much more efficiently. Let’s use the Mariners as an example of what I’m talking about.
For 2005 through 2007, the Mariners are going to end up having spent around $270 million on their player salaries. They essentially spent it in equal chunks, with about $85 million going out in 2005, $88 million in 2006, and a projected $95 million or so in 2006.
If they had shifted to a win budget system after the debacle of 2004 and correctly analyzed the talent they had on hand, they could have fielded a 65-70 win team in 2005 for about $70 million. Last offseason, the goal was to get to 80-85 wins, which could have been accomplished on a payroll of about $80 million.
That would leave them with $120 million dollars, not including the accrued interest gained from the money not spent in prior years, to spend on the 2006 Mariners, the year that they’ve been mandated to win or lose their jobs. Would you like the M’s to have $120 million payroll next year? Yea, me too.
A win budget system systematically shifts payroll dollars away from years where the team is rebuilding and into years where the team is challenging for the World Series. It maximizes the impact of the wins the teams are paying for, and acknowledges that a 95th win in one season is worth more than the 72nd win in another season.
Someday, someone in a front office is going to convince their owners to shift to a win budget system. And their economic advantage is going to be staggering.
At first glance this is actually what I thought Toronto was doing over the last few seasons. But it turns out they were just spending more money because they could.
Brilliant post.
You could do away with the luxury tax (at least in the NL) and probably revenue sharing too….since everyone is shooting for 84 wins, with this system, there would be little danger of anyone in the NL having a $100M payroll…. 😛
Why exactly is it that teams are so tied to the idea of a fixed budget? I’m not very familiar with the corporate accounting process and I suspect that on some financial level there’s a good reason that teams don’t vary their payroll from year-to-year. Clearly, there’s the “message to the fan” level of the situation where teams don’t want to make fans think they are waving the white flag, but there must be more to it than that.
Someday, someone in a front office is going to convince their owners to shift to a win budget system. And their economic advantage is going to be staggering.
and even after that, it is going to then take a good long while for the glacial movements of change to hit anyone else’s front office, especially the one closest to home.
and I think we need a general off-season thread, for breaking news like Gerald Perry to the Cubs as hitting coach, Bud Black leaving the Angels for SD and [flash!] the new D’Backs
uniforms
and re: budgeting, Phil Rogers puts his two cents in on the pricey offseason to be…
Dave,
It seems this idea has a lot of potential as a tool to *think about* putting a team together, but actually running a team this way might be difficult. There are two assumptions here that are problematic:
1: “In a win-budget concept, a team would do a rational, honest evaluation of how many games the current roster could be expected to win if each additional hole was filled with a replacement level player.” I don’t think I need to tell you how difficult that is. How many wins would such an evaluation have predicted for the 2004 M’s? How about 2005? I’d posit that both would’ve missed the mark by over 20%. I can see why the business owners might not be happy with that level of uncertainty. This isn’t an issue with the evaluations being too rose-tinted, it’s just that there’s a whole lot of noise in these numbers, and PECOTA/Marcel/ZIPS still haven’t fixed that problem (they’re clearly better than the alternatives).
2: The assumption that the ‘wins’ available in the marketplace are basically the same from year to year; that wins are wins, and you can always go to the market and buy some. In your thought experiment, the M’s don’t go after Beltre and Sexson in the rebuilding phase, but ‘bank’ the cash to purchase wins now. Sounds good, but outside of Matsuzaka, I’m not sure a substantially higher payroll helps the M’s *this year*. Schmidt? Zito? They could take on lots of salary in trade, but I’m not sure they’ve got the pieces to get, say, Miggy Cabrera. The M’s spent when they did both because they apparently misjudged how many wins they had on the roster in 2004, AND because they liked those players. Beltre hasn’t exactly worked out, but I can’t say I fault their logic.
So how do we work around those obstacles? Is there a better way to gauge where your team currently stands? Is there a way to ensure that the wins you get actually show up on the field?
#3, Mat, I think for teams that don’t have a lot of guaranteed revenue, the payroll *does* shift substantially. The M’s may be somewhat atypical. I think the rich teams don’t vary it, especially downwards, because they simply don’t have to. The M’s really thought they’d compete in 2005 (and 2004); the idea that they could’ve gotten 70 wins for $65-70m would’ve seemed like a non-sequitor. And that brings us back to how difficult it is to accurately gauge how many wins you’ve got on your team….
I suspect that on some financial level there’s a good reason
Well…it’d be nice to believe there is good reason. However my years in the corporate world tells me that financial people simply like to eliminate as many varibles as possible, cut any expense that accountants don’t understand/personally value and having carbon copy budgets makes projections less risky for personal careers.
The corporate world survives mostly on the backs of hard-working, risk taking employees who will shun the process to accomplish the goal. Some companies flourish despite executive management, rarely due to management.
rant off/
I’m not really disgruntled however…
I’m also curious about how this would work, since free agents rarely sign one-year contracts. Obviously the year-to-year evaluation would have to take place with a certain amount of extra forecasting on what the team this new player would be a part of would be capable of doing, and who might be available to place around that player and how much he would add or detract from the future win values of that team. You have to forecast that player’s value in the future years of the contract, too. Predicting a player’s performance for the upcoming season is hard enough. Projecting it for two or three years beyond that is next to impossible, with all the variables that would be involved.
I’m not sure I’m explaining my thought well…
I don’t think I need to tell you how difficult that is. How many wins would such an evaluation have predicted for the 2004 M’s?
2004 was probably inevitable, even though Derek, Jason, and I all wrote extensively about how poorly that team was put together, and that the organization was clearly headed in the wrong direction. I’m not suggesting that going to a win budget concept would always lead teams to performing up to expectations. I am suggesting, however, that teams could be significantly more realistic with their expectations, and more efficient in how they construct their roster to reach those goals.
The assumption that the ‘wins’ available in the marketplace are basically the same from year to year; that wins are wins, and you can always go to the market and buy some.
Believe it or not, this is true. The marketplace isn’t just free agency, despite the fact that fans continually use the terms interchangeably. There are always wins to be acquired, and there are always good ways to spend your money.
In your thought experiment, the M’s don’t go after Beltre and Sexson in the rebuilding phase, but ‘bank’ the cash to purchase wins now.
In my experiment, the M’s don’t go after Sexson, but they do acquire Beltre. A win budget philosophy is not avoidance of acquiring star talent in down years – it’s simply a shift towards acquiring talent that complements your team’s needs at the time. The Mariners needed a 26-year-old franchise third baseman, but they did not need a 29-year-old mediocre first baseman.
So how do we work around those obstacles? Is there a better way to gauge where your team currently stands? Is there a way to ensure that the wins you get actually show up on the field?
Hire smarter people.
You have to forecast that player’s value in the future years of the contract, too. Predicting a player’s performance for the upcoming season is hard enough. Projecting it for two or three years beyond that is next to impossible, with all the variables that would be involved.
Any team that is not already forecasting a player’s value in the future years of a contract doesn’t deserve to win. And I disagree with how hard projections are as a group – on any given player, there is wide variance, but easily understandable principles can be drawn from historical data and adhered to with good success.
It’s not that tough to say that signing 34-year-old pitchers with declining strikeout rates to four year contracts is a bad idea. The variability of forecasting in a small sample isn’t as relevant when looking a team building philosophy. And that’s really what we’re talking about here.
I’m not disagreeing that specific trends can most definitely be applied, and that you can come up with generalalities based on historical precedent. But since there are a reasonable amount of outliers, and a team may have a specific need that has to be filled, with a limited supply of players available to fill that need, it becomes trickier to balance the general with the specific. You have to base decisions on risk, and there are certainly better risks to take than others. Beltre, of course, from most appearances, was a fairly safe bet (many different things aligned very well to point to him being a reasonable risk). In his specific 2005, I don’t think anyone projected he’d perform as poorly as he did.
It’s applying generalities to specific players that becomes quite challenging. I’m not sure that can be ignored in this process.
It’s applying generalities to specific players that becomes quite challenging. I’m not sure that can be ignored in this process.
The best teams operate under established philosophies that they believe in, and don’t go chasing after exceptions to the rule. The A’s, Indians, Braves, and Twins are perfect examples of this – they have pretty strict organizational beliefs, and you won’t see them trying to figure out if Free Agent X is the outlier that goes against what they believe.
Not applying generalities to specific players is a great way to become the 2004-2006 Mariners. Good teams operate in the margins of probability. Bad teams believe that probabilities are too often wrong to put stock in.
One potential problem with this notion, Dave, is that a team can run into a heckuva problem when they’re on a downcycle. Say a team looks at its situation rationally and says, “We’re budgeting for 70 wins this year,” and that leaks to the media.
The outcry would be substantial. And a story that juicy has to leak to the media, doesn’t it?
But teams already do that all the time. They don’t say “70 wins” but they say “rebuilding” or set a payroll that’s not competitive.
Really to me the key is being will to shift payroll across years. I.e., if you save $15M this year by trading players or not signing free agents, you have the money to spend “next” year. The Mariners don’t/won’t do that.
But if you argue that “And I disagree with how hard projections are as a group – on any given player, there is wide variance, but easily understandable principles can be drawn from historical data and adhered to with good success,” then Paul’s right – this isn’t a way to get to an actual, on-the-field 25-man roster, it’s a philosophical exercise.
If the noise is too great on individual players, then how does that help you in the market where the wins you’re buying are attached to actual, specific ballplayers?
“2004 was probably inevitable, even though Derek, Jason, and I all wrote extensively about how poorly that team was put together, and that the organization was clearly headed in the wrong direction.”
You’re absolutely right, and you recognized this before basically all the fans. But you still predicted, as I hazily recall, something north of 80 wins. When even the smart guys can be wrong by 20% (and that’s not uncommon), we might need a new model.
None of this is to say that the current system is totally rad, or that incremental budgeting is awesome. I’m just struggling to see how this would work in real terms. What would a smart GM using this model do differently from a smart GM NOT using this model?
With all of the discussion about building a roster creatively in what seems to be an outrageous free agent market, Cleveland landed Josh Barfield today to plug their 2B hole. Obviously the M’s have no need for up the middle youth, however this should be a wake-up to other GMs looking to upgrade by signing a Carlos Lee type disaster.
Barfield goes to Cleveland
# 13
The success cycle is pretty common in sports and fans understand that sometimes, their teams need to rebuild. You could combat the public backlash by hiring good PR people to get people excited about the team’s “young stars in the making”. When a team is rebuilding, more often than not, they are building around cheap young talent that they hope can become stars. In these years, the payroll would naturally be lower than in years when they are paying primo bucks to top players. Where a team can get themselves into trouble is by not being honest with where they are in the success cycle.
A good GM has to sell a realistic plan to ownership that may include some rebuilding. Unfortunately, pro sports is such a high stakes, competitive game that some GMs are not given the latitude to rebuild when they need to.
#6 Marc: Totally agree. The main problem presently is that teams (the vast majority and nobody has a crystal ball…) do a bad job of evaluating both what they have and what they’re getting. They need to ‘hire smarter people’ presently. This proposition doesn’t address those issues.
I completely agree with using this model as a guideline of figuring out what quality of addition you need to plug into whatever holes your team has. It doesn’t change the uncertainty of health and decline. At least most of the main problems inherent in the current model are still inherent with this. You still haveto rely on evaluation and predictions. You still haveto factor in having $57.2 million dollars (random figure)already spent for the next season before you can figure where you’re at. Obviously there’s a huge difference between a team you figure to be worth 70 wins on paper in a high revenue situation with only $50 million in gauranteed contracs and a team you figure to be worth 70 wins, low revenue with already $35 million in contracts. That’s without looking at how many holes you have to fill. Too many other factors to simplify the whole process.
Evaluation, prediction, where to spend for upgrades in wins (production), how to split up the overall combined wins ($$$)
You’re still looking for bargains. The main difference I see is the ability to save money for a couple years (which then can create a losing philosophy which is a whole other problem) and then spend a little more over the next few years. The only other changes I see are verbage, and but maybe I’m missing something.
Incidentally I’d love to see the M’s not spend so much this offseason and shift the difference to 08 and 09 on the FA class of 2008. If Bavasi figures out how to spend every penny of next years budget this year (in multiyear contracts) there will be a lot of unhappy fans in Seattle a year from now. I can’t imagine looking at those particuliar names that address longtime needs of this team knowing that the M’s haveto trade someone away to have the finances available to get them. Best FA class in more than 10 years…EASILY!
The main difference I see is the ability to save money for a couple years (which then can create a losing philosophy which is a whole other problem) and then spend a little more over the next few years.
So do we expect now that Miguel Cabrera can’t lead a team to a World Series because the Marlins spent about $15M in payroll last year? I don’t think there’s much of a need to worry about a losing attitude developing. Even teams with low projections are generally going to have high spots during the year that make them think they can be contenders. I’d guess that’s enough to keep them away from a losing attitude.
It’s a lot more difficult to assemble a team talented enough to win the World Series than it is to motivate players to want to win the World Series, so I think teams should spend most of their efforts fixing the former problem.
The success cycle is pretty common in sports and fans understand that sometimes, their teams need to rebuild.
The success cycle is, and always has been, a myth. I’ve debunked this at BP. Teams don’t inevitably decline if they’re good, they don’t inevitably get better if they’re bad, they don’t get better if they’re average after they were bad, they don’t get worse if they’re average after they were good.
Teams are what they are. Success cycles are pattern recognition where there is no pattern.
DMZ – What do you think of “The Plexiglass Principle”?
I think that marc w. (#6, 15) and PositivePaul (#8) summed up my reaction to this. In fact, my first thought was something akin to what Paul articulated. There are problems with this because the market for free agents (and yes, Dave, I know that’s not the only way to skin the cat) isn’t fluid because players don’t all sign one year deals. [To some extent, I’ve had some of the same doubts about “freely available” replacement-level talent, too, though I think that does exist, though maybe not quite as much as assumed in stats circles.] I also think marc’s point about the level of noise in evaluating wins, on an individual level, complicates the task.
On the whole, though, this is a framework for thinking about talent acquisition and roster-building that is intriguing and should be pursued, even if it won’t work perfectly. The system teams operate under now doesn’t work 90%+ of the time – it would certainly be an improvement.
I like the idea as a much better overall approach, but one problem I see is that with such a flexible budget, in the year you want to spend that extra $30 million to get to 95-100 wins, you might have to lock yourself in to extending that extra money into seasons where you would prefer a lower budget. It’s probably harder to add those extra wins on one-year contracts. Am I missing something?
The only thing that I would add is that, from a business perspective, this is quite a simple concept. Multiply the marginal win by the expected revenue increase it will generate and compare that to the best option in terms of marginal wins/salary. Compare those two and the decision is easy as to whether it makes sense to spend more from a financial perspective. You’d want to be careful to include any opportunity costs (e.g. we’ve got money tied up in Sexson and this inhibits our ability to acquire new players, but it also means we don’t have to spend money on a first baseman). This is really just a generalization of Dave’s budgeting wins concept.
It’s not quite so simple in reality as there are lag effects (as all M’s fans should be well aware by this point). However, this can be accounted for by discounting the future expected values and summing up, giving one number (in a multi-year contract, for instance). The only thing left then is risk. A ten year contract is probably not a good idea even if the expected marginal win per dollar looks good. I guess I just assumed this was the type of analysis teams did. Upon further reflection, I would agree that it’s not exactly what they do. The annual dollar budget is a silly concept.
With respect to opportunity cost, I would suggest that it is a bad idea to acquire much in this FA market as it will inhibit the acquiring teams’ ability to get talent in the future, when the FA market is likely to be better. That is, the opportunity cost of acquiring FA’s this year is high (salaries will be high and talent will be low, relatively speaking).
Very interesting discussion….I thought this is what the A’s essentially do….have a budget, but acquire players focusing on how they expect their team to perform. For example they traded off both Hudson and Mulder a year before they would hit free agency, and used the players acquired in those trades to position themselves to contend in future years. 2005 was a “rebuilding” year in which some of their young players got the necessary experience, and Billy Beane put a club together in 2006 that took the division, and advanced one step into the playoffs. Please recall how he signed Loaiza at the very beginning of the offseason, and that filled a hole in their rotation, and ultimately lead to the Division crown.
The best teams operate under established philosophies that they believe in, and don’t go chasing after exceptions to the rule.
It’s pretty clear the Mariners do this, too, by their “we will bid X dollars over Y years for PLayer Z, because that’s what we think he’s worth” approach to free agency.
The problem is the established philosophy they use nets them Richie Sexson, Carl Everett and Jarrod Washburn. Whoops.
Derek,
You may find this post interesting, which essentially tells you how many wins you should have, if you spend to your market size:
http://www.tangotiger.net/archives/stud0060.html
This post here tells you how many wins you expect based on actual payroll, with post #9 being the most up-to-date of the data:
http://www.insidethebook.com/ee/index.php/site/comments/competing_on_payroll/
I think you mean “Dave”
Oops, sorry about that.
[ot!]
[ot]
[ot]
[ot]
[ot]
So, were those last 5 posts off topic? I’m just curious.
We are fans. We want the team to win. The team is a business. Its owners want it to be profitable. The problem for the fans is that there are several strategies for a team owner to maximize profit (given constraints like existing contracts and stadium), and only one of them (at most) involves actually playing in the postseason or even having a winning season (for other strategies, see Kansas City, Florida in the years following its WS wins…and for that matter, the Cubs or the Red Sox for many years of the past century).
It would seem that the Cardinals are ahead of the curve. MGL says in this post:
The “trick” for an owner of GM (a fiscally responsible one) is to be able to estimate their team’s chances of making the playoffs in any year and at any time (in-season, off-season, etc.) in order that they are able to quantify that marginal win value in $, based on their increased or decreased chances of qualifying for the post-season.
One of my responsibilities with the Cardinals is to send the owner a report which estimates every team’s expected w/l record and chances of arriving at every point in the post-season. For example, last year we were prohibitive favorites (90 something percent) to make the post-season and therefore presumably out of that so-called sweet spot. It probably would have made financial sense to either try and dump a little salary and actually worsen the team or at least not to spend any more money trying to make the team better.
One of the mistakes that aggressive owners (and GM’s I suppose) make is spending money to strengthen their team with the goal of “winning the world series” as opposed to just making the post-season. For example, had we (STL) spent more money last year to try and improve our chances of winning a WS title (by maybe getting another front-line pitcher) it probably would have amounted to a negative ROI. Although the BP book chapter did not look at how much money a pennant or WS win was independently worth to a team, as opposed to just a post-season appearance, even if it were a lot, it is difficult to increase a team’s chances of winning the WS by more than a few percentage points.
Why can’t we get guys like that? (With apologies to Boston sportscaster Bob Lobel…)
This is essentially constant-dollar-averaging investment strategy, just applied to the ups and downs of a baseball team rather than a stock. (If the M’s were a stock, I’d have been short selling for quite a while now…)